Narrow SCSI - Expectations of Target's
Ashok Raj
ashokr at pyramid.com
Thu Oct 9 15:23:16 PDT 1997
* From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
* ashokr at pyramid.com (Ashok Raj)
*
[ Henry Wong wrote...]
>
> * From the T10 (formerly SCSI) Reflector (t10 at symbios.com), posted by:
> * Henry Wong <hwong at asic.qntm.com>
> *
> Being in ASICs I seldom get out of the 1's & 0's world.
> I would like to hear what is the correct or prevelant way
> to handle the following 'system' scenario (which is
> out of my league and handled by the firmware folks).
>
> I would appreciate any & all feedback from both host
> adapter & target device folks!!
I have not completely read this.. but from what i understand
looks like you are thinking that narrow is during selection phase.
but i beleive that during selection phase all ids are visible on
the wide bus. being narrow on pon is only for data phases i think.
pls correct me if iam wrong.
ashokr
>
> BASIC SCENARIO:
> 1) A 'physically' WIDE Initiator with an ID=15
> 2) A 'physically' WIDE Target with an ID=1
> 3) System does not have SCAM & ID's are
> hard jumpered...let's keep this simple
> 5) SCSI System is WIDE with Narrow only
> devices mixed in ..but the Target doesn't
> know this yet.
> 6) Target defaults to Narrow mode (let's
> say after power-up) since it doesn't
> know if the Bus if physically Narrow
> or Wide yet
> 7) The Initiator trys to select the Target
>
> ==> Now which of the three cases below is "correct"?
> => Case1: The target is only monitoring the lower
> SCSI Bus bits since it has reverted to
> Narrow and is not aware if the System is
> a Wide or Narrow Bus. It responds as if
> it "thinks" it's in a SASI system because
> it see's only one ID (it's own ID) on the
> Bus. Yes.. the target knows it's not a
> real SASI system since it gets the Identify
> /ATN phase. The target basically handles
> the command w/o Disconnect. After Cmd
> cmplt, switches to Wide (until negotiated
> otherwise).
> => Case2: The target being physically Wide always
> monitors the high SCSI Bus bits even if
> it has reverted to Narrow mode...thus
> will respond to the Initiator as if
> it were Wide and business as usual as
> a "psuedo-wide device"...versus.. a
> Narrow device in a "real" Narrow bus.
> => Case3: Have the target ignore the Selection
> all together since it doesn't know if
> it must assume it is Narrow because it
> may be on a Narrow Bus and that it should
> not expect to be seeing a SASI system.
> Also, knowing it won't know if it is a
> SASI system until it responded to the
> Selection.
>
> Perhaps "the" basic question is should the Target (knowing
> it's SCSI connector is physically Wide... ie. 68-pin
> or 80-pin SCA-2) always observe and react accordingly
> to what is present on the High SCSI Bus byte. In
> the case of a Selection, always monitor bits 15-8+parity
> for proper ID & good parity before responding <irrespective
> of whether or not the Target is defaulting to Narrow
> Bus transfer mode>?? This also implies that the SCSI
> system MUST terminate the High Bus if it intends to use
> and adapt a physically Wide device for a Narrow-only
> SCSI system. Also, the High Bus Byte is ONLY ignored
> when the Target PCB is physically Narrow (ie 50-pin
> connectored)??
>
> Thanks!
> Regards, Henry
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at symbios.com
>
--
============================================================================
-m------- E S Ashok Raj e-mail : ashokr at pyramid.com
---mmm----- SMTS I/O Software Group Phone (W) : (503)-690-6216
-----mmmmm--- Siemens Pyramid Fax(W) : (503)-690-7704
-------mmmmmmm- Information Systems, Inc
__>
__/o \-----\ @ Think Pig... o__,
|__ PIG \ The Pyramid I/O Group _,>/ _
\________/ (*)`\(*)
_|| _||
============================================================================
*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at symbios.com
More information about the T10
mailing list